Abstract
This study attempted to explore the role of group motivation (i.e. collective efficacy) and collaborative learning behavior (i.e. group cohesion; group cognitive involvement) in students' group performance in learning computer science in computer-supported collaborative learning environment. 96 Taiwanese college students (32 groups) enrolled in the course "Web Page Programming" participated in the study. The students discussed and processed their group projects using "Intelligent Collaborative Learning System". The results indicated that collective efficacy predict group cohesion significantly in CSCL. In addition, collective efficacy also showed to predict group performance in programming significantly. Moreover, the results also indicated that both group cohesion and groups with higher cognitive involvement in their discussion leaded to better group performance. In summary, this study validated the social cognitive model in learning computer science in CSCL, that is, groups with higher collective efficacy had stronger group cohesion and better performance in learning Web Page programming. Groups with higher cognitive involvement and group cohesion also leaded to better group performance.